Conditional parameter fitting with arbitrary objective function

[jongkeesbj]

Currently, conditional parameter fitting in the PEC can only be done when passing a column in the data argument of the PEC and referring to this column in the depends_on argument. This is fine if the aim is to fit parameters to data, for example here.

However, we do not always pass a data argument, for example when the aim is to optimize a user-defined objective function (such as reward rate). See an example here.

Essentially, what is desired is the option to pass a column to the PEC that can be referred to in the depends_on argument without relying on the data argument. Passing anything to the data argument prevents manually specifying an objective function, which means it is impossible to conditionally fit parameters using a user-defined objective function.

Edit: to clarify, I think there are two separate issues: 1) the ability to directly pass a column/data frame to the PEC which can be referred to in the depends_on argument, and 2) the fact that passing anything to data prevents also passing anything to objective_function. My current request pertains to the first issue.